In recent years, real-world external controls (ECs) have grown in popularity as a tool to empower randomized placebo-controlled trials (RPCTs), particularly in rare diseases or cases where balanced randomization is unethical or impractical. However, as ECs are not always comparable to the RPCTs, direct borrowing ECs without scrutiny may heavily bias the treatment effect estimator. Our paper proposes a data-adaptive integrative framework capable of preventing unknown biases of ECs. The adaptive nature is achieved by dynamically sorting out a set of comparable ECs via bias penalization. Our proposed method can simultaneously achieve (a) the semiparametric efficiency bound when the ECs are comparable and (b) selective borrowing that mitigates the impact of the existence of incomparable ECs. Furthermore, we establish statistical guarantees, including consistency, asymptotic distribution, and inference, providing type-I error control and good power. Extensive simulations and two real-data applications show that the proposed method leads to improved performance over the RPCT-only estimator across various bias-generating scenarios.

翻译：近年来，真实世界外部对照（ECs）作为增强随机安慰剂对照试验（RPCTs）能力的工具日益流行，尤其在罕见病或平衡随机化不符合伦理或不切实际的情况下。然而，由于ECs与RPCTs并非总是可比，不经审慎直接借用ECs可能会严重偏倚治疗效应估计量。本文提出一种数据自适应整合框架，能够通过偏倚惩罚动态筛选出可比ECs来实现自适应特征。该方法可同时实现：（a）当ECs可比时达到半参数效率界；（b）通过选择性借用减轻不可比ECs存在的影响。此外，我们建立了包含一致性、渐近分布和推断在内的统计保证，提供I类错误控制和良好检验功效。大量模拟实验和两项真实数据应用表明，在各种偏倚产生情境下，所提方法相比仅使用RPCT的估计量能带来更优性能。